Inventory and pricing management in probabilistic selling

Résumé

Context: Probabilistic selling is the strategy that the seller creates an additional probabilistic product using existing products. The exact information is unknown to customers until they receive the probabilistic products. This strategy is still a relatively new area for both researchers and practitioners. Many of the corresponding operations problems need to be solved to take full advantage of the opportunity of this innovative marketing strategy. However, limited attention has been paid to examining the inventory management of probabilistic selling from the perspective of Operations Management, which cannot meet the needs of decision-making in reality. Objectives: Considering different characteristics of the probabilistic product, the buyer, and the seller involved in probabilistic selling, i.e., the probabilistic product form, the buyers’ behaviours of demand switch and barter exchange, and the seller's product allocation behaviour, we establish models and solve the decision problems of pricing, inventory, joint decision of pricing-inventory, and product allocation, etc. Based on the analysis of optimal decisions and strategy comparison results, we shed some lights on the effectiveness of probabilistic selling on managing uncertainty, and its profitability. Method: First, we analyze the practice scenarios of probabilistic selling. Next we mainly use newsvendor inventory model, hotelling model, and optimization theory to model, solve, and analyze the operational problems. Then we give some analytical results. Next we conduct the numerical analysis using softwares of Matlab and Mathematica. Finally, we provide insightful managerial implications for the practice of probabilistic selling. Results: The thesis derives the optimal operational decisions of inventory order, pricing, inventory allocation, and product line design in probabilistic selling. Overall, the analysis of the results show that probabilistic selling can benefit the seller with higher expected profit by reducing demand/supply uncertainty and improving inventory efficiency. The performance of probabilistic selling is closely dependent on customers' price sensitivity, product similarity, and uncertainty level, etc. Main results considering different research scenarios are as follows: 1) When the price for the probabilistic product is independent on demand reshape, a proper cannibalization can benefit the retailer in terms of yielding a higher expected profit. Probabilistic selling is more profitable with relatively lower product similarity and higher price-sensitive customers, while inventory substitution strategy outperforms probabilistic selling with higher product similarity. 2) When the price for the probabilistic product is dependent on demand reshape, probabilistic selling can benefit the seller with higher expected profit and lower inventory. Probabilistic selling is more profitable with lower product differentiation, higher customers' price sensitivity, and higher demand uncertainty. Improper pricing would undermine the seller's profit. 3) When the seller offers physical probabilistic product, he can benefit from two effects, namely the risk pooling effect due to demand reshape and the risk diversification effect due to inventory flexibility. 4) When the seller offers barter choice in probabilistic selling, he may benefit from the marketing effect in the barter process. Offering barter choice can broaden the application range of probabilistic selling, which will increase with successful barter probability. Conclusions/Implications: First, the thesis helps sellers understand how to manage their inventory, pricing and related implementation issues to take full advantage of probabilistic selling. Second, this thesis explores the mechanism of this innovative marketing strategy as an inventory management tool to combat uncertainty which also riches the literature on Operations Management, especially inventory management.